Introduction

Following the seminal works by Sherrington on the organization of the Rolandic cortex using focal electrical stimulation (Sherrington 1906); Penfield and Boldrey (1937) demonstrated a somatotopic organization of the sensory-motor system in humans. The authors performed intraoperative electrostimulations of the motor and somatosensory cortex in awake patients to demonstrate the existence of a so-called “homonculus”. In practice, they used these findings to tailor the surgical resection according to the results of intraoperative functional mapping. More recently, direct electrical stimulation of the pyramidal pathways also showed a somatotopy of the corticospinal tracts, with induction of involuntary movements of the lower limb, upper limb, and face by stimulating from mesial to lateral directions (Duffau et al. 2003).

Moreover, Penfield and Jasper also described cessation of voluntary muscle contraction when selected areas rostral to the primary motor cortex were electrically stimulated (Penfield 1954). Lüders et al. confirmed these observations and localized the sites within the posterior part of the inferior frontal gyrus as well as the pre-supplementary motor area (SMA). They termed the identified locations ‘‘negative motor areas’’(Lüders et al. 1987, 1995). Additional negative motor areas have been identified anterior to the primary motor cortex, corresponding to the premotor cortex and the SMA-proper (Mikuni et al. 2006). Such negative motor responses have been observed for the controlateral leg, arm, and face (Filevich et al. 2012; Lüders et al. 1995; Schucht et al. 2013), as well as for bimanual task (Rech et al. 2014) and speech (Kinoshita et al. 2015). These negative phenomena are defined as a complete arrest of movement or speech without loss of consciousness (Lüders et al. 1987). However, recent stimulation studies evidenced more complex interferences with motor function, resulting not only in inhibition but also in acceleration of movement, supporting the existence of a network devoted to motor control (Schucht et al. 2013).

Interestingly, a somatotopy of this circuit underpinning movement control has been suggested at the level of the cortex (Godschalk et al. 1995; Picard and Strick 1996), the basal ganglia (Nambu 2011), and the spinal cord (Dum and Strick 1996; He et al. 1993)—especially in animals thanks to the use of microstimulation. Nonetheless, so far, no findings have been reported concerning a possible somatotopic organization of the white matter tracts involved in motor control in humans. Of note, in previous studies in which we used intraoperative electrical stimulation mapping, we observed the existence of subcortical stimulation sites generating cessation of movements, with a veil-like distribution, anterior to the primary motor fibers: therefore, we suggested the existence of descending pathways originating from premotor areas—areas which are known for negative motor response characteristics (Rech et al. 2014; Schucht et al. 2013). Here, using the same methodology, we investigated whether this motor control network driving negative motor responses had a somatotopic distribution.

To this end, in a series of eighteen patients who underwent awake surgery for a frontal low-grade glioma, we assessed the negative motor responses and their distribution throughout the white matter located under premotor areas. The goal of surgery was to improve the onco-functional balance by maximizing the extent of resection while preserving motor skills and thus quality of life (De Benedictis et al. 2010; Duffau and Mandonnet 2013). Intraoperatively, all patients experienced cessation of the movement of lower and upper limb, of coordinated bimanual movements, and/or speech. It is worth noting that these subcortical sites were somatotopically distributed, with arrest of movement of the lower limb, upper limb(s), and face/speech by stimulating from mesial to lateral directions, and from posterior to anterior directions. Therefore, we describe for the first time, to our knowledge, a somatotopic organization of the white matter bundles underpinning motor control in humans.

Materials and methods

Patient population

We report a prospective series of 18 patients with a frontal World Health Organization grade II glioma. They were selected for surgical resection from May 2012 to May 2014 in our department. Due to the proximity of the tumor with motor structures, all patients underwent awake surgery with intraoperative cortical and subcortical electrostimulation mapping of motor functions.

A neurological examination as well as a standard neuropsychological and language assessment was performed before and few days after surgery. Informed consent was obtained prior to surgery from all patients.

Intraoperative electrostimulation mapping protocol

Resection was performed under local anesthesia so that functional cortical and subcortical mapping could be carried out using direct brain stimulation, as already described in previous studies (Duffau et al. 2003; Duffau 2005, 2015). A bipolar electrode (tip-to-tip distance: 5 mm) delivering a biphasic current (pulse frequency of 60 Hz, single pulse phase duration of 1 ms, amplitude from 1 to 4 mA (Nimbus*) was applied on the brain.

First, ultrasonography was used to identify the tumor boundaries. Then, the cortical mapping was performed over the primary sensory-motor area (without motor task) and ventral premotor cortex (during a counting and naming task) by progressively increasing the level of stimulation of 0.5 mA (from a baseline of 1 mA) until a functional response (involuntary movement, dysesthesia, anarthria or naming disturbances) was elicited—indicating the optimal threshold of stimulation. All positive stimulation sites were marked with a tag number.

During a second surgical stage, the removal of the tumor was performed according to the cortical and subcortical functional boundaries identified with electrical stimulation throughout the resection. The same electrical parameters were used at the cortical and subcortical as well.

During surgery, patients were asked to perform a language and a motor task. More specifically, these tasks consisted of repetitive and alternative flexion and extension of the controlateral arm, hand, fingers, and lower limb at approximately 0.5 Hz (i.e., one flexion/extension cycle every 2 s), combined with a naming task (composed of 80 items)—the latter task allowing to map not only the different language processes (i.e., semantic, phonological, and syntactic) but also motor-related speech processes such as articulation and face movements (Tate et al. 2014). Of note, although the patient is awake in the operating theater during 1–2 h throughout the mapping and tumor resection, which implies a strong constraint for the selection of tasks, such repetitive movements are feasible by the patients, as already reported (Fernández Coello et al. 2013; Kinoshita et al. 2015; Rech et al. 2014; Schucht et al. 2013). A well-experienced neuropsychologist (GH) and speech therapist (SMG) analyzed the functional disturbances elicited by electrostimulation.

A negative motor response (NMR) was defined as cessation of the movement without loss of consciousness (Lüders et al. 1987, 1995). In a first step, the goal was to observe whether stimulation elicited NMRs in the controlateral hemibody. Then, the patient was asked to perform bimanual movements, that is, a series of bilateral flexion and extension of the two arms, hands, and fingers, in phase or in anti-phase, at a frequency at approximately 0.5 Hz. In all cases, the neuropsychologist/speech therapist checked whether the movements (i) were made continuously or whether they stopped, (ii) whether there was a modification of the frequency (e.g., acceleration or slowdown), and (iii), whether there was a modification of the bilateral coordination (e.g., in-phase movement which became anti-phase or vice versa). The stimulation was first done at the same subcortical sites which previously elicited controlateral NMR, and then around this area, that is, on the posterior part of the surgical cavity for frontal glioma. An NMR was defined as bilateral (BNMR) when synchronous cessation of bimanual movement occurred without loss of consciousness. In all cases, movement arrest was observed by external visual inspection without any cinematic or neurophysiologic quantification.

The corresponding subcortical sites were marked with a tag number in case of reproducible movement interference, namely when clinical symptoms were induced by at least three stimulations. An intraoperative photography was taken to show the final cortico-subcortical eloquent sites. Of note, the resection was extended up to functional boundaries assessed by stimulation mapping that preserved both crucial cortical areas and essential subcortical pathways—to optimize the extent of resection while preserving brain functions (Duffau 2009, 2012).

The neurological status was assessed immediately after surgery and again after 3 months, analogical to the preoperative assessment.

Normalization procedure and replacement of stimulations sites

For each patient, a high-resolution 3DT1 (resolution: isometric voxels—1 mm) was acquired 3 months after surgery. This imaging sequence was then normalized in the MNI space using cost function masking (Brett et al. 2001). Briefly, the resection cavity was first manually drawn using MRIcron software (http://www.mccauslandcenter.sc.edu/mricro) and binarized to constitute a mask. This mask was used during the registration process carried out with SPM 8 (http://www.fil.ion.ucl.ac.uk/spm/software/spm8/) implemented in the Matlab environment (http://fr.mathworks.com/products/matlab/). Then the resection cavities were drawn again to yield an individual volume of interest (VOI).

The exact location of the stimulation site was determined on the normalized scans using various anatomical (gyri, sulci, midline, deep gray nuclei, and lateral ventricle) and functional (motor, sensory, and language) landmarks, as documented on intraoperative photography. Indeed, eloquent areas are localized in the periphery of the cavity as the resection was stopped according to functional boundaries. This cavity can be seen on the postoperative MRI; thus the correlation between imaging and postoperative photography can be done. The stimulation sites were then plotted on the MNI 152 template. A 1-mm diameter spherical VOI was created for each stimulation site and the corresponding MNI coordinates were registered (see Online resource 1 and 2).

To better analyze the spatial distribution of the stimulations sites, it was mandatory to replace the VOIs on the same side by taking the absolute value of the coordinate for each stimulation site. For the illustration purpose, the stimulation sites were plotted into a 3D rendering (the MNI 152 template) using MRIcroGL (http://www.mccauslandcenter.sc.edu/mricrogl/). To visualize the spatial distribution of the stimulations sites, in an illustrative case, the resection cavity has been shown by removing the brain overlaid with the normalized VOI in the MNI152 template.

Given the number of stimulation sites and the different types of responses, spatial distribution of NMRs was analyzed separately, by overlapping each kind of NMR two by two (face/upper limb—upper limb/BNMR—upper limb/lower limb) into a 3D rendering (MNI 152 template). This allowed a better analysis of the spatial relation between each kind of NMR.

In the last analysis, the mean coordinates of each kind of responses were obtained after switching each coordinate in its absolute value. Results were then plotted on a 3D rendering.

Results

Patients

The mean age of patients (9 males) was 31.9 years (range: 27–65). 15 patients were right-handed, 2 were left-handed and 1 was ambidextrous according to the Edinburgh inventory (Oldfield 1971). None of them had motor deficit (especially no motor initiation disturbance) or language impairment before surgery.

According to the preoperative anatomical MRI, the tumor involved the frontal lobe in the 18 cases (10 right, 8 left) within or at least close to one of the following motor-related structures: dorsal premotor cortex (dPMC), SMA, anterior cingulate cortex, head of the caudate nucleus or anterior arm of the internal capsule (Table 1).

Table 1 Patient characteristics and follow-up
Full size table

Characteristics of cortico-subcortical stimulation sites

Positive motor responses were elicited for each patient over the precentral gyrus, corresponding to the primary motor cortex (M1) (Fig. 1). These positive responses were distributed somatotopically over M1. They consisted of involuntary muscle contraction of the controlateral limb or hemi-face. As previously mentioned, due to the limitation of time and clinical constraints, our protocol of cortical mapping did not include continuous movements of the extremities, and, therefore, cortical NMRs were not recorded.

Fig. 1
figure 1

Illustrative case (patient 3): a Postoperative MRI showing the resection cavity. M1 corresponds to the primary motor cortex. b Three-dimensional view of the resection cavity. Brain has been rotated to correspond to the operative view. c Intraoperative view after brain mapping at the end of the resection. 1, 2, speech arrest corresponding to the ventral premotor cortex; 3, Primary motor cortex of the face generating face contraction; 4, 5, 6, Primary motor cortex of upper limb generating an involuntary movement of the hand and arm; 7, dorsolateral prefrontal cortex generating language perseverations. At a subcortical level: 47 (under the vein), stimulations generated NMR of the face; 49, NMR of the face and the upper limb; 50, BNMR; 48, NMR of the upper and lower limbs. Correspondence between the three-dimensional and the intraoperative views are shown with color lines and numbers

Full size image

However, at the subcortical level, NMRs were elicited for each patient (Figs. 1, 2). NMRs involving face or articulation elicited speech arrest (Fig. 2a). Those involving limbs consisted of an immediate and complete cessation of arm, hand, fingers (Fig. 2b), and/or lower limb (Fig. 2d) movement without any loss of consciousness and without any loss of tonus (namely, the limbs did not fall at the moment of stimulation). Bilateral NMR (BNMR), defined as synchronous and NMR of both upper limbs during a bimanual task, was also generated during subcortical stimulation (Fig. 2c).

Fig. 2
figure 2

Repartition of sites eliciting NMR according to the kind of NMR observed. All sites have been reported on a single side for a better visual appreciation. Each dot corresponds to a site of NMR (n = 18 patients). Axial, coronal and sagittal views (with and without ipsilateral brain) are shown on each row. Each column corresponds to a single kind of NMR. First column (a) with yellow dots: face/speech NMR. Second column (b) with blue dots: controlateral upper limb NMR. Third column (c) with white dots: BNMR. Fourth column (d) with red dots: lower limb NMR. Spatial distribution between each kind of NMR can, therefore, be analyzed for each plane (see Figs. 4, 5, 6 and 7 for an overlapping of each type of NMR overlapping)

Full size image

Sites of stimulation were located at the level of the white matter underneath the dPMC and posterior part of the SMA, immediately in front of the precentral sulcus—rostral to the corticospinal tracts, whatever the hemisphere considered (Figs. 2, 3; Online resource 1).

Fig. 3
figure 3

Overlapping of sites eliciting NMR on a single template with axial (a), frontal (b) and sagittal (c) views. Each dot corresponds to a site of NMRs (n = 18 patients). The same colors than those in Fig. 3 have been used. Sites eliciting two NMRs are shown in different colors. Yellow, face/speech NMR; Blue, controlateral upper limb NMR; Gray, face/speech and upper limb NMR; White, BNMR; Light blue, BNMR and upper limb NMR; Red, lower limb NMR; Violet, upper and lower limb NMR. M1, primary motor cortex; PreCG, precentral gyrus (for separate views allowing comparison between the two kinds of responses, see Figs. 4, 5, 6 and 7)

Full size image

Interestingly, each type of NMR (namely speech/upper limb/BNMR/lower limb) was elicited in a specific location within the white matter, with a systematic and reproducible distribution (Fig. 3). From Figs. 3, 4, 5, 6, and 7, each kind of NMR is represented by a specific colored dot. When the stimulations led to combined NMRs (see below), a combination of the two colors has been used. Figures 4, 5, and 6 show the relationships between each kind of specific NMR by overlapping them two by two (i.e., face/upper limb, upper limb/BNMR, upper/lower limbs) to better understand the organization of the subcortical NMR sites. Finally, Fig. 7 shows the average spatial distribution of the NMR site for each kind of NMR to allow a better understanding of the organization of the subcortical NMR sites. In a coronal plane, laterally, in the white matter located under the posterior part of the inferior frontal gyrus—in front of the ventral premotor cortex (VPM), that is, the lateral part of the precentral gyrus (van Geemen et al. 2014), stimulation elicited speech arrest (Fig. 4). Slightly more medially, in the white matter underneath the dPMC—in front of the hand knob—stimulation generated NMR involving the controlateral upper limb (Fig. 4). In 11 patients, BMNRs were induced during stimulation performed between the sites eliciting unilateral NMRs of the upper limb, in the same coronal plane (Fig. 5). Medially, by stimulating the white matter underneath the posterior part of the SMA-proper, in front of the paracentral lobule, NMRs involving the lower limb were elicited (Fig. 6). Furthermore, in a sagittal plane, NMRs were disposed in an antero-posterior direction, with sites generating face/speech arrest rostrally located, then sites eliciting upper limb NMRs more posteriorly located and finally, a more caudal location for sites inducing lower limb NMRs (Figs. 4, 6, 7).

Fig. 4
figure 4

Overlapping of sites eliciting NMR of the face/speech and controlateral upper limb (see the legend of Fig. 3). This figure shows the lateral, ventral and anterior repartition of sites eliciting NMR for face/speech (yellow), whereas upper limb NMR (blue) are localized more medially, rostrally and posteriorly

Full size image
Fig. 5
figure 5

Overlapping of sites eliciting NMR of controlateral upper limb and BNMR (see the legend of Fig. 3). This figure shows that the sites eliciting BNMR (white) are localized within those eliciting controlateral upper limb NMR (blue)

Full size image
Fig. 6
figure 6

Overlapping of sites eliciting NMR of controlateral upper and lower limbs (see the legend of Fig. 3). This figure shows that lower limb NMR (red) are localized more posteriorly, medially and dorsally than upper limb NMR (blue)

Full size image
Fig. 7
figure 7

The mean coordinates for each kind of NMR have been computed, as detailed in the “Materials and methods” section. These average coordinates have been then reported on the MNI template. Yellow, face NMR (mean ± SD MNI coordinates: x = 19 ± 5, y = 5 ± 8, z = 34 ± 10); Blue, upper limb NMR (MNI coordinates: x = 16 ± 6, y = 4 ± 8, z = 39 ± 12); White, BNMR (x = 15 ± 4, y = 1 ± 10, z = 44 ± 9); Red, lower limb (x = 11 ± 3, y = 0 ± 8 z = 45 ± 8)

Full size image

In addition, single stimulation sometimes led to combined NMRs, as cessation of both lower and upper limbs by stimulating the white matter at the junction of the sites corresponding to the upper limb NMR and lower limb NMR, respectively—or as speech arrest associated with upper limb NMR by stimulating the white matter at the junction of the sites corresponding to speech arrest and upper limb NMR, respectively.

Subcortical fibers responsible for NMRs were followed deeper throughout the resection until the anterior arm of the internal capsule and the head of the caudate nucleus.

Postoperative course

All patients recovered well from surgery and were discharged home within 5 days following surgery. Four patients experienced a slight paresis of the upper limb, 4 had a worsening of the verbal fluency, and 3 a mutism whose one with a complete akinesia of the hemibody. All patients with neurological disorders underwent rehabilitation at home. On re-examination at 3 months, all patients had regained their respective preoperative level, with no motor neither speech deficits.

Discussion

Spatial distribution of the NMR sites

Since the description of the Penfieldian homunculus, the somatotopic organization of M1 is well established (Penfield and Boldrey 1937; Rizzolatti et al. 1998). This spatial distribution exists not only at the cortical level, but also subcortically. Indeed, anatomic dissections have demonstrated that the pyramidal pathways are somatotopically organized in the corona radiata, with, from medial to lateral directions, white matter fibers of the lower limb, then of the upper limb, and then of the face (Déjerine 1895). More recently, this somatotopy has also been confirmed using intraoperative electrical stimulation of the white matter tracts (Duffau et al. 2003). This distribution also exists in the internal capsule, the cerebral peduncle, and the spinal cord (Dum and Strick 2002; Kwon et al. 2011).

Based upon this concept, several studies have reported that other cortices had a similar somatotopic organization, as the premotor areas (Dum and Strick 1991; Godschalk et al. 1995; Picard and Strick 1996) and the SMA (Cauda et al. 2011; Fontaine et al. 2002). Interestingly, these regions seem to participate in a circuit involved in motor control, because negative motor phenomena with cessation of movement or speech arrest have been observed when these structures are electrically stimulated (Lüders et al. 1987, 1995; Penfield 1954). Indeed, during stimulation, cortical sites eliciting speech arrest/face NMR are more ventrally and laterally located, whereas cortical sites generating NMR of the upper limb are more dorsally and medially situated, in front of knob of the hand (Enatsu et al. 2013; Mikuni et al. 2006). Nevertheless, despite a better understanding of this premotor cortical distribution, the organization of the white matter tracts subserving such a motor control network has received less attention.

Recently, we showed that direct subcortical electrostimulation of descending fibers, anterior to the corticospinal tracts and originating from premotor cortex, were also able to induce suppression of unilateral or bilateral movements (Rech et al. 2014; Schucht et al. 2013). Further, stimulation sites in the parietal white matter as well as in the anterior arm of the internal capsule indicated a large-scale fronto-parietal motor control network (Almairac et al. 2014). Here, using the same methodology of intraoperative subcortical stimulation in awake patients operated for a premotor low-grade glioma, we confirmed our previous results, since we generated NMRs of the upper and lower limbs, the face, speech arrest, and BNMRs.

Moreover, our present study brings original insights into the organization of the white matter subserving movement control, by demonstrating a somatotopic distribution of the subcortical sites eliciting NMRs. Indeed, in a coronal plane, from medial to lateral, stimulation elicited NMRs of the lower limb, then of the upper limb(s), and then of the face with speech arrest (Figs. 4, 5, 6, 7). In addition, in a sagittal plane, NMRs were disposed in an antero-posterior direction, with sites generating face/speech arrest rostrally located, then sites eliciting upper limb NMRs more posteriorly located and finally, a more caudal location for sites inducing lower limb NMRs (Figs. 4, 5, 6, 7). Another argument supporting a somatotopic organization of this network is the fact that a single stimulation may elicit combined NMRs, as cessation of both lower and upper limbs by stimulating the white matter at the junction of the sites corresponding to the upper limb NMR and lower limb NMR, respectively—or as speech arrest associated with upper limb NMR by stimulating the more lateral white matter at the junction of the sites corresponding to speech arrest and upper limb NMR, respectively. Such a phenomenon is likely due to combined stimulation of different bundles when converging in the depth (Fig. 3). However, another explanation is that cessation of both lower and upper limbs could be due to stimulation of fibers that simultaneously control upper and lower limbs motor activity—and that might be crucial for actions as walk or run. Finally, within the fibers controlling the upper limbs, we found a smaller zone eliciting BNMR and underpinning the bilateral motor control pathway that we have already described (Rech et al. 2014) (Fig. 5). This spatial distribution was the same in each patient, whatever his/her handedness and whatever the side.

It is worth noting that there is a coherency between the somatotopic organization of the premotor cortex described in previous reports [i.e., cortical sites eliciting speech arrest/face NMR more rostrally and laterally located, and cortical sites generating NMR of the upper limb more dorsally and medially situated (Enatsu et al. 2013; Mikuni et al. 2006)] and the somatotopy of the white matter tracts underneath these cortical areas, as defined by electrical stimulation. This pleads in favor of the fact that these fibers are descending from the “negative motor cortex” involved in the control of movement in humans. Furthermore, the spatial distribution of the white matter bundles underpinning the motor control network is similar to the somatotopic organization of the pyramidal tracts. Thus, we could hypothesize that each part of M1 (and its associated corticospinal fibers), corresponding to the representation of a specific part of the body according to the homonculus, might interact with a dedicated portion of the cortico-subcortical network involved in the control of movement. This seems anatomically plausible through the short frontal U-fibers connecting M1 with the premotor cortex (Catani et al. 2012).

White matter tracts underlying the motor control network

We can only speculate about the tracts mediating this network, because no tractography has been performed in this study. Indeed, our goal here was to use the unique opportunity provided by subcortical stimulation to study the functional distribution of the fibers underpinning the motor control circuit, in particular by investigating the existence of a potential somatotopy. Nonetheless, we have already assumed in previous studies that the subcortical fibers responsible for NMRs/BNMRs seemed to take their origin in the premotor cortex and in the posterior part of the SMA-proper and that they run to the head of the caudate nucleus as well as to the internal capsule (Rech et al. 2014; Schucht et al. 2013). Interestingly, by combining diffusion tensor imaging and intraoperative subcortical mapping, we have recently demonstrated that stimulation of the fronto-striatal tract (FST, which connects the SMA with the anterior part of the caudate) generated NMRs (Kinoshita et al. 2015). Therefore, since both SMA and caudate are known to be involved in the initiation and control of movement (Alexander and Crutcher 1990; Swann et al. 2012), we hypothesized that the FST could be part of the motor control network. Thanks to the new findings reported in our present study, and knowing that anatomically the FST is more medially located, we can suggest that this tract might be particularly involved in the control of lower and upper limb movements. In addition, we have also previously described that direct stimulation of the frontal aslant tract (FAT, which connects the pre-SMA with the inferior frontal gyrus [Catani et al. 2012]) elicited speech disturbances (Kinoshita et al. 2015). This is in agreement with the fact that lesions within the anterior part of the SMA can induce disorders of speech initiation—with sometimes a complete mutism (Krainik et al. 2003). In the same vein, the inferior frontal gyrus is thought to be implicated in speech programming (Amunts et al. 2004). As a consequence, on the basis of our original data reported here, we may propose that the FAT, running more laterally than the FST, could be more specifically involved in the control of face and speech. Furthermore, in the sagittal plane, the FAT is more anteriorly located than the FST (Kinoshita et al. 2015), which is in line with the somatotopy of the SMA, i.e., with the representations of language, the face, the upper limb, and the lower limb located from anterior to posterior (Fontaine et al. 2002). Remarkably, this is also in agreement with our observations using subcortical stimulation, which showed a more rostral location of NMR sites for face/speech and a more caudal location of NMR/BNMR sites for limbs (Figs. 4, 6 and 7). Thus, we hypothesize that the motor control network is a complex circuit constituted by multiple tracts, including U-fibers (as suggested above), associative fibers (FAT), and projection fibers (FST), somatotopically organized. Moreover, the bilateral organization of this wide multi-bundle network, as supported by induction of NMRs/BNMRs during stimulation of both hemispheres, might explain plasticity mechanisms underlying functional improvement after a so-called “SMA-syndrome” [with transient disorders of movement initiation following SMA damage (Krainik et al. 2004)] (see below).

Functional role of the “motor control pathway”

Many hypotheses have been formulated concerning its functional significance, and even if still matter of debate (Filevich et al. 2012), its implication in the control of (complex) movement seems to be very likely (Enatsu et al. 2013; Rech et al. 2014; Schucht et al. 2013). Indeed, although Krainik et al. showed that resection of the SMA-proper could lead to postoperative permanent deficit in bimanual coordination and fine motor skills (Krainik et al. 2001), Schucht et al. reported that intrasurgical preservation of this large cortico-subcortical motor control network avoided definitive motor disorders (Schucht et al. 2013). Rech et al. confirmed these results and introduced the concept of a bilateral motor control pathway involved in bimanual coordination (Rech et al. 2014). In the present series, thanks to the fact that NMR/BNMR sites have been detected and preserved during surgery, all our patients completely recovered—even though transitory hemiparesis and speech disturbances have been observed in the immediate postoperative period, because the resection was pursued up to negative motor structures. Taken together, these results support the crucial role of this circuit in motor control, based not only on initiation but also on inhibition of movement.

In this setting, it seems logical to benefit from a somatotopical organization of this large-scale multi-bundle network, to better coordinate the four limbs and/or speech in complex movements. Indeed, one may speculate that each limb should be independently supervised (even if part of a global action) when requiring interlimb coordination, as playing piano/organ (eventually by singing simultaneously), or dancing, or playing sports, etc. Beyond intraoperative electrostimulation mapping, further studies are needed to improve our understanding of the mechanisms mediating such real-time high-order control in ecological conditions.

Limitations of this study

To optimize glioma resection while preserving quality of life, electrical stimulation was performed to identify functional boundaries and was, therefore, restricted to certain areas. As a consequence, we should acknowledge that further stimulation sites leading to various motor or speech responses might have been missed; especially, other cortical areas and fibers involved in this complex network may have been ignored.

Conclusions

This is the first evidence of a somatotopic organization (both in coronal and sagittal planes) of the white matter bundles underpinning movement control in humans. A better knowledge of the distribution of this motor control network may have important implications both in cognitive neurosciences and clinical practice, especially in brain surgery.